Do I need all that LLM reasoning?An experiment with GPT-6 Luna on 240 puzzles.Nov 9, 2026 · 10 min read · 7 views
